Add auto-advance: click Next, scroll to top, capture next page
After each capture the controller asks the vision model to locate a 'Next' control (returned as normalized screen coordinates), clicks it via OS input, scrolls to top, and continues the loop. Configurable via SCREEN_LEADS_AUTO_NEXT / _NEXT_LOAD_PAUSE / _MAX_AUTO_NEXT, with a per-run safety cap. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
@@ -147,6 +147,9 @@ python main.py</code></pre>
|
||||
<tr><td><code>ANTHROPIC_API_KEY</code></td><td>—</td><td>Required. Claude API key.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_MODEL</code></td><td>claude-opus-4-8</td><td>Vision model (use claude-sonnet-5 to cut cost).</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_START_DELAY</code></td><td>5</td><td>Grace seconds before first capture (0 = off).</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_AUTO_NEXT</code></td><td>true</td><td>After each capture, click a "Next" control, scroll to top, and continue.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_NEXT_LOAD_PAUSE</code></td><td>2.5</td><td>Seconds to wait for the next page to load after clicking Next.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_MAX_AUTO_NEXT</code></td><td>25</td><td>Safety cap on auto-advances per run.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_SCROLL_STEPS</code></td><td>6</td><td>Screenshots per profile.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_SCROLL_AMOUNT</code></td><td>800</td><td>Scroll distance per step.</td></tr>
|
||||
<tr><td><code>SCREEN_LEADS_SCROLL_PAUSE</code></td><td>0.8</td><td>Settle time (s) after each scroll.</td></tr>
|
||||
|
||||
Reference in New Issue
Block a user